I am using the TPS62153 as a pre-regulator that steps +12V DC down to +5V DC as an input for a secondary buck regulator which makes dual 3.3V supplies. The design appears to work as expected, but after a day or so of nominal operation 62153 goes into an unrecoverable latch-up. In the latch-up condition, the part pulls anywhere around 0.3A from 12V input, all of which appears to be power dissipated by and within the 62153 itself. Nothing I have tried so far (power cycling, cooling, etc.) has successfully un-latched it once it gets into this condition, so it appears to be permanent damage.
The seeming randomness of when this strikes has made it difficult to determine what's triggering it. At first I thought this was likely due to voltage over-stress caused by a bench supply that had some particularly nasty power-on transients, but that has long been remedied and the yet latch-up problem still persists. This has now happened on three different boards (and more than once on one of the boards that was re-worked to replace the 62153), so clearly there is a problem lurking somewhere in the design.
I've been over the schematic design, data sheet, layout, etc. several times now but haven't yet spotted the culprit. Perhaps this rings a bell with someone on the forum? I attached schematic with the relevant circuits.
Thanks in advance.LED CTRL.pdf
